Adaptive energy management strategy for fuel cell/battery hybrid vehicles using Pontryagin's Minimal Principle

The fuel cell hybrid vehicle provides an efficient and low-emission alternative for the internal combustion engine vehicle. The energy management strategy (EMS) commands the power split between the power sources and is crucial to the hybrid vehicles. In this work, we propose an adaptive EMS based on Pontryagin's Minimal Principle for a fuel cell/battery hybrid vehicle, in which the co-state adaptation is performed by driving cycle prediction. In order to improve the co-state estimation accuracy, an improved Markov based velocity prediction is proposed considering the driving behavior under different driving patterns. Moreover, the driving pattern is recognized online based on a support vector machine method, which is optimized by particle swarm optimization. We build a combined driving cycle to verify the effectiveness of the proposed strategy, simulation results under three cases show that the proposed strategy can foresee the driving behaviors and update the co-state reasonably. Comparing with the rule-based EMS, the proposed strategy achieves a better fuel economy with a hydrogen consumption reduction by 4% and a relatively low average power change rate of fuel cells. Moreover, it achieves a close performance to the offline optimal algorithms.